The incidence of adenocarcinoma (AC) of the esophagus has increased rapidly in most countries during the past three decades, yet the reasons are not well understood. AC typically arises on a background of Barrett's esophagus (BE). At the population level, relatively little is known about the environmental and genetic causes of BE and AC, or about factors which modify the natural history of BE to cause progression to cancer. In this population-based study, we aim to quantify the risks associated with exposure to epidemiologic and genetic risk factors for reflux esophagitis (RE), BE and AC. In parallel biospecimen analyses, we aim to identify molecular subtypes of BE and AC using microarray gene expression profiling and tissue arrays. To accomplish these aims, we will sample representative groups of patients with biopsyproven RE [n=400], BE [n=700] or AC [n=300] from all pathology laboratories servicing the target populations during a 3 year period, and compare them with two groups of controls. A representative group of population controls [n=600] will be sampled from a compulsory electoral register, and a group of biopsynegative tissue controls [n=400] will be sampled from the pathology laboratories. Cases and controls will answer identical questionnaires, focusing on gastro-intestinal symptoms, exposure to medications (especially reflux promoters, NSAIDs, COX-2 inhibitors, hormones), as well as smoking, alcohol, infection with Helicobacter pylori and family history of cancer. Blood samples will be collected from participants to identify genotypes associated with predisposition to RE, BE and AC. From cases and tissue controls, we will obtain specimens of biopsy or surgical tissue with the aim of determining the prevalence of molecular subtypes of BE and AC. Fresh tissue will be available from a proportion of clinic-based AC cases [n=50-100] for gene discovery through microarray gene expression profiling. From comprehensive mining of the expression profiling data we will identify diagnostic markers for the different subtypes of BE and AC, prognostic markers that predict likelihood of progression and/or response to therapy, and targets for rational drug design to treat BE and AC. Candidate genes will be validated in the paraffin sections available for all cases and tissue controls using tissue array technology. Epidemiologic analyses will then be performed comparing risks of exposure among the major disease groupings (RE, BE, AC), as well as for the molecular subtypes of RE, BE and AC.